Politecnico di Milano

Master Thesis

Simulation and Optimization of an oven

hinge mechanism through
LMS-AMESim

Advisor:
Prof. Monica Bordegoni

Author:

Supervisor:

Milad Amini Harandi

Eng. Francesco Ferrise

Scuola di ingegneria industriale e dellinformazione

Dipartimento di Meccanica

September 2013

I remember my mothers prayers and they have always followed me. They have clung
to me all my life.

Abraham Lincoln

POLITECNICO DI MILANO

Abstract
Dipartimento di Meccanica
Master Of Science
Simulation and Optimization of an oven hinge mechanism through
LMS-AMESim
by Milad Amini Harandi

This thesis describes the simulation and optimization of an oven hinge by means of the
software LMS AMESim. The real hinge model helps us to select the elements of different
libraries of the software namely signal, mechanical, planar mechanic and cam
and follower besides, calculating the needed parameters for each icon.
After running the simulation the CAD models are imported in AMESim to have a
realization of the assembly. Then an experiment is performed to discover the input
load applied to the handle of the oven and find out about the values of velocity and
acceleration after running the simulation. Optimization of the system is done based on
minimizing the value of the integral between the squared difference of the velocity of the
real model and the system. After performing the optimization via genetic algorithm
and NLPQL method, the corresponding value of all implemented elements are obtained.
Finally modeling of the power grasp (holding the handle of the oven) is done by an eight
element circuit. The results of this section are compared with the experiment.

Questa tesi descrive la simulazione e ottimizzazione di una cerniera forno mediante il

software LMS AMESim. Il modello di cerniera vera ci aiuta a selezionare gli elementi
di erenti biblioteche del software e cio segnale , meccanico, meccanico planare
e cam e seguace . Inoltre, il calcolo dei parametri necessari per ogni icona. Dopo
aver eseguito la simulazione dei modelli CAD sono importati in AMESim per avere un
realizzazione del montaggio. Poi viene eseguito un esperimento per scoprire lingresso
carico applicato alla maniglia del forno e nd conoscere i valori di velocit e accelerazione
dopo lesecuzione della simulazione. Ottimizzazione del sistema viene fatto sulla base
minimizzando il valore dellintegrale tra il quadrato di erenza della velocit del vero
modello e il sistema. Dopo aver eseguito lottimizzazione tramite algoritmo genetico e
metodo NLPQL, il corrispondente valore di tutti gli elementi implementati sono ottenuti.
Infine modellazione della presa di potenza (tenendo la maniglia del forno) fatto da un
otto circuito elemento. I risultati di questa sezione sono confrontati con lesperimento.

Acknowledgements
I would like to express my gratitude to my dear supervisors Francesco Ferrise and Serena
Graziosi for the useful comments, remarks and engagement through the learning process
of this master thesis. Furthermore I would like to thank Prof. Monica Bordegoni for
introducing me to the topic as well for the support on the way. Last but not least, I like
to thank my parents for their endless love and support.

iv

Contents
Abstract

ii

Acknowledgements

iv

List of Figures

vi

List of Tables

viii

1 Introduction
1.1 Introduction . . . . . . . . . . . . .
1.2 Product Design and Developement
1.3 Industrial Design . . . . . . . . . .
1.4 Solutions . . . . . . . . . . . . . .
1.4.1 LMS AMESim . . . . . . .

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

2 Simulation
2.1 Case Study . . . . . . . . . . . . . . . . . . .
2.1.1 Computer Aided Design Drawings . .
2.2 AMESim . . . . . . . . . . . . . . . . . . . .
2.3 Opening-closing mechanism of the door of the

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

1
1
1
3
5
10

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

11
11
12
18
19

3 Optimization
3.1 Experiment . . . . . . . . . . . . . . . . . . . . . . .
3.2 Optimization and parameterization of the case study
3.2.1 Export . . . . . . . . . . . . . . . . . . . . . .
3.3 A Model of the User as a Mechanical Load . . . . .
3.3.1 Model . . . . . . . . . . . . . . . . . . . . . .

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

37
37
41
43
50
50

. . .
. . .
. . .
oven

4 Conclusion

58

A References

60

List of Figures
1.1
1.2
1.3
1.4
1.5
1.6

An iPod, an industrially designed product. . . . . . . . . . . .

Typical design process from Lawson [5] . . . . . . . . . . . . .
An example of 2D drawing . . . . . . . . . . . . . . . . . . .
Finite element analysis of a component belonging to TON die
A scale model of an airplane in a wind tunnel for testing. . .
Complete computer simulation of a bacterium . . . . . . . . .

. . . . .
. . . . .
. . . . .
casting
. . . . .
. . . . .

.
.
.
.
.
.

. 4
. 5
. 6
. 7
. 8
. 10

2.1
2.2
2.3
2.4
2.5
2.6
2.7
2.8
2.9
2.10
2.11
2.12
2.13
2.14
2.15
2.16
2.17
2.18
2.19
2.20
2.21
2.22
2.23
2.24
2.25
2.26
2.27
2.28
2.29
2.30
2.31

LUCE SX 1046LPX S . . . . . . . . . . . . . . . .
Features . . . . . . . . . . . . . . . . . . . . . . . .
Technical details . . . . . . . . . . . . . . . . . . .
Housing . . . . . . . . . . . . . . . . . . . . . . . .
Shells to be placed on the back of the oven . . . .
Side panels . . . . . . . . . . . . . . . . . . . . . .
Gasket . . . . . . . . . . . . . . . . . . . . . . . . .
Door compartments . . . . . . . . . . . . . . . . .
Door frame . . . . . . . . . . . . . . . . . . . . . .
Diplomat hinge . . . . . . . . . . . . . . . . . . . .
The location of hinge . . . . . . . . . . . . . . . . .
AMESim Simulink lab . . . . . . . . . . . . . . . .
Opening-closing mechanism of the door of the oven
Main body block diagram . . . . . . . . . . . . . .
Port positions in parameter mode . . . . . . . . . .
Real hinge . . . . . . . . . . . . . . . . . . . . . . .
Force directions . . . . . . . . . . . . . . . . . . . .
First part of block diagrams . . . . . . . . . . . . .
Second part of block diagrams . . . . . . . . . . . .
Third part of block diagrams . . . . . . . . . . . .
Fourth part of block diagrams . . . . . . . . . . . .
Set parameters for cam and follower mechanism . .
Cam profile shown with light blue line . . . . . . .
Cam profile in 360 . . . . . . . . . . . . . . . . . .
Fifth part of block diagrams . . . . . . . . . . . . .
Sixth section Block diagrams . . . . . . . . . . . .
AMESim parameter . . . . . . . . . . . . . . . . .
Run Parameters . . . . . . . . . . . . . . . . . . .
Start a simulation . . . . . . . . . . . . . . . . . .
Assembly mode . . . . . . . . . . . . . . . . . . . .
Assembly mode . . . . . . . . . . . . . . . . . . . .

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

vi

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

11
12
12
13
14
15
15
16
17
17
18
19
20
21
21
21
22
23
25
26
27
27
28
28
30
31
31
32
32
33
33

List of Figures
2.32 Animation of the opening-closing mechanism of the door of the oven with
Large opening . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.33 Stay-rod variable list . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.34 Graph for position, velocity and acceleration of the main body . . . . .
2.35 Animation of the opening-closing mechanism of the door of the oven with
small opening . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.36 Graph for position, velocity and acceleration of the main body with snap
back effect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.1
3.2
3.3
3.4
3.5
3.6
3.7
3.8
3.9
3.10
3.11
3.12
3.13
3.14
3.15
3.16
3.17
3.18
3.19
3.20
3.21
3.22
3.23
3.24
3.25
3.26
3.27
3.28
3.29
3.30
3.31

vii

. 33
. 34
. 34
. 35
. 36

An overview of the sensors installed on the oven . . . . . . . . . . . . . .

Load sensor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
PCB Pieszotronics accelerometer . . . . . . . . . . . . . . . . . . . . . . .
Single Axis Watson VSG . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Initial and final position of the door of the oven . . . . . . . . . . . . . . .
Experiment data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Curve fitting of load data vs. time . . . . . . . . . . . . . . . . . . . . . .
Curve fitting of acceleration data vs. time . . . . . . . . . . . . . . . . . .
Curve fitting of velocity data vs. time . . . . . . . . . . . . . . . . . . . .
New load function f (x) = 356.8 exp(((x 3.691)/0.07581)2 ) . . . . . .
New load function f (x) = 356.8 exp(((x 3.691)/0.07581)2 ) . . . . . .
New load function f (x) = y exp(((x3.777)/0.1385)2 )+z exp(((x
3.68)/0.07037)2 )) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Export setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Export setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Input parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Simple output parameters . . . . . . . . . . . . . . . . . . . . . . . . . . .
Compound output parameters . . . . . . . . . . . . . . . . . . . . . . . . .
Simple output parameters . . . . . . . . . . . . . . . . . . . . . . . . . . .
Design exploration dialogue box . . . . . . . . . . . . . . . . . . . . . . . .
Genetic Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
First optimization based on the Genetic Algorithm . . . . . . . . . . . . .
Solution progress dialogue box . . . . . . . . . . . . . . . . . . . . . . . .
Optimized velocity of the model vs. Real velocity obtained by measurement
Eight-element model of the users impedance, modeling the passive mechanics for frequencies >20 Hz. . . . . . . . . . . . . . . . . . . . . . . . .
Global parameter setup dialog box . . . . . . . . . . . . . . . . . . . . . .
Hand user elements implemented in AMESim . . . . . . . . . . . . . . . .
Hand user elements implemented in main model . . . . . . . . . . . . . .
The mechanical imedance function in parameter mode . . . . . . . . . . .
Mean values of the mechanical impedance model according to figure 3.24
for different grasping siutations . . . . . . . . . . . . . . . . . . . . . . . .
Impedance measurement settings for different grasps [13] . . . . . . . . . .
The plot of the force exchanged between finger-tip, bones and muscles . .

38
38
39
39
40
40
41
41
42
42
42
43
44
45
45
46
47
47
47
48
48
49
50
51
53
53
55
56
56
56
57

List of Tables
3.1

Best solution found via the criterion minimizing final value of the integral
of squared difference between speed of the modeled system and that of
the real experiment. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49

viii

All my love to my mom, for finding me the light, whenever it was

far away. . .

ix

Chapter 1

Introduction
1.1

Introduction

In 21st century, the competitive market pushes the companies to launch their products
faster and cheaper. So in order to decrease the time of product design process, new
methodologies suggesting Virtual prototyping method is introduced in the first chapter.
The advantages and effects of this method are explained briefly in the following.
In the second chapter the case study which is an oven distributed by Indesit Company
is dismounted part by part. The software LMS AMESim is introduced to simulate the
opening-closing mechanism of the door of the oven.
In the third chapter the optimization is clarified step by step thanks to the software
AMESim. Finally modeling of the user as a mechanical load is carried out.

1.2

Product Design and Developement

The definition of the term product includes a range of simple objects starting from a
book or a bottle of water to intricate objects like home appliances or even electronic
circuit boards. The word Design means a set of plans and drawings showing how a
product is to be made in sense of geometry, material or styling and generally it deals
with the function and the look of the product. Design could be simple or complex such as
fitting the chain of a bicycle at the proper place or building IC chips respectively. Tong
and Sriram [1] describe design as a process that constructs a description of an artifact,
process, or instrument that satisfies a (possibly informal) functional specification, meets
certain performance criteria and resource limitations, is realizable, and satisfies criteria
such as simplicity, testability, manufacturability, and reusability.
1

Chapter 1. Introduction

Development deals with product marketing meaning identifying the market opportunities and trying to build and represent what is demanded to be sold out to customers.
Companies work with clients every time and it is important that they can find the right
product to begin with. Because if they do not start with a suitable product, it will be
harder to revise it further at the succeeding industrial design stages.
Now a question arises. What is a right product? Firstly, it has to satisfy the correspondent demand sent from market. Actually those demands need to be straight forward
meaning simple and clear. So the product design process is the transformation of an idea,
a demand from the market place or a need of a purchaser to the design of a product that
satisfies that need. The stages are explained here briefly to make it easier to understand.
Besides, additional terms in marketing are introduced here shortly.
Fuzzy Front End
The term Fuzzy front end or what is abbreviated as FFE simply refers to the stage where
designers get together in order to figure out what a product should be, what a product
should do or even what a customer experience should be. At this level, they do research
activities like creating focus groups or conducting voice of the customer interviews
where they learn how to define a product. In other words, the organizations decide if it
is worth to invest resources on an idea at further development stages or not. This term
became famous by Smith and Reinertsen [2] in 1991 and is counted as one of the early
stages of product design process and roughly covers the period from generating an idea
to its satisfactory for development or elimination. The effort made in this stage has a
good effect on the whole innovation process.
Voice of the customer and pain points
Voice of the customer or simply VOC is tightly associated with marketing and identifying
the lack of comfortability of any item, like when a customer declares that a particular
product does not fill in his/her needs in a second way. Thus when we are looking for
pain points, in fact we are seeking an opportunity to create a solution. In other words
it can be considered as an interaction between the customers and the marketers. The
marketers listen to positive and especially negative opinions of their customers via many
channels such as groups, communities, blogs and etc. Nowadays companies, brands and
campaigns overcome the lack of responding towards their purchasers. They compensate
it with quickness in answering and getting feedback from their products end-user which
could drive the companies forward like a motor in a system.

Chapter 1. Introduction

VOC trends
The trends come through a variety of different industrial groups where different colors,
different materials and different textures are all combined to create interest from the
customer and of a particular buying pattern. VOC trend is approved by the industry
leaders and project managers as a next great business battlefield. According to Aberdeen
group [3] associated with Harte-Hanks Inc. in USA, the customers would like to see an
increased attention on the product services and support from their companies despite a
fast decision for resolving defects in the products.
VOC relation to Form and Fuction
Nowadays, companies should take into account the customers ideas and opinions and
link it with all the decisions made in the whole systems-engineering field. The voice of
the customer must be used in parallel in the development level.
Companies always try to make a balance with form and function, depending on the
company itself on which factor spends more time but generally the inputs for function is
gained from voice of the customer particularly pain points and for form designers should
care about the leading trends that influence how a product looks acceptable.

1.3

Industrial Design

The role of industrial design or ID is getting broader and broader everyday where sometimes intersects with marketing and engineering. We take marketing as an input, like
who the customer is, what the customer wants, what comes back from customer studies, where the market opportunities are and etc. Instead industrial designers face with
designing products. ID is an original role and certainly is one branch of design that has
much in common with architecture and with many other fields such as engineering.
Industrial designers take input from the engineering and marketing and then doing
research in depth to explore the customer interests. They start creating a solution
around the function and form and merry those two in a way that not only aesthetic but
also an experience that is considered as a product solution.
Product solution concept starts off with rough sketches, moving through concepts rendering, typically through industrial design process, they go back to the customer that
the sketches and ideas and renderings make sense with the customer.
Once the renderings are confirmed then we move into the 3D computer aided design
drawing simply called CAD drawings where we actually have a delivery board there.

Chapter 1. Introduction

Figure 1.1: An iPod, an industrially designed product.

Up to this stage, there has been no interacting in an increased level with engineering
department to make sure that the design, the choice of materials and the geometries
make sense except through the 3D CAD design stage that the design engineering have
the most communication where the industrial designers make sure that the exterior
function make sense while the engineering have negotiation with industrial designer
about the process to get assured that they optimize the whole process in order to make
a successful product.
According to industrial designers, a product has to include inherently two main characteristics including a great looking product and great functioning. Primarily accomplished their task then it turns into defending that aesthetic and keep it up through the
rest of the ramp up to manufacturing. This is controlled by checking and making sure
that their changes into the model are due to the manufacturing and it is done in a way
that maintains the original design that was created through the core of industrial design
process.
Products are designed to make an awesome first impression on the customers so it is
important to take packaging and branding into account in order to create a lasting
emotional impact to the customer.
typical design strategy As explained before the design process is based on an analysis
following design strategies and processes. This process is about how the industrial
designers face with the problems. Asimov [4] demonstrated it as a sequence of analysis,
synthesis and evaluation inside each stage generalized to the whole process Figure 1.2.
There is a list of typical strategies in the following:

Chapter 1. Introduction

Figure 1.2: Typical design process from Lawson [5]

1. Feasibility Study
2. Preliminary Design
3. Detailed Design
4. Planning the production process
5. Planning for distribution
6. Planning for consumption
7. Planning for the retirement of the products

Typical design process

1. Analysis
2. Synthesis
3. Evaluation - extended to
4. Revision
5. Implementation

1.4

Solutions

When somebody has a product to process we typically start out with some kind of 3D or
2D representation of the geometry either from a computer aided design namely CAD or
sometimes we can start with 2D drawings namely sketches fig. 1.3 and we start building
3D representation in CAD. In fact CAD profits from the advantage that all the drawings
are in form of standard sheets and are used in every communications. Thus CAD is 3D

Chapter 1. Introduction

Figure 1.3: An example of 2D drawing

computer aided design system that allows up to shape or to form of a product. We can
also check fitting a product during assembly to see how everything is located and check
the tolerances. Also CAD is used to re-design and optimize a product.
Finite Element Analysis
The nest step in the process of finding a solution is the simulation portion. That is
usually done in Finite Element Analysis or what is known as FEA software. This is
the stage where we take a continuous model either developed in a CAD system or in
the finite element software itself. The solid model or 2d representation (the sketch)
are discretized by using meshing in the finite element software in order to be analyzed
structurally, thermally , magnetically and etc in a single simulation using best-in-class
software to predict and optimize the performance.
Parametric Design Quite often in simulation and design or optimization there are
parameters that you might want to vary to see what the effects would be. It is called
parametric design and it can be done in FEM software as well as the CAD software. It
might be the height, thickness, width of a part or even diameter of a cylinder. It could
be considered as a parameter so that we can analyze, simulate and optimize the design
by considering all those aspects of design to get the best performance with the lowest
cost.

Chapter 1. Introduction

Figure 1.4: Finite element analysis of a component belonging to TON die casting

Prototyping
Prototyping is defined as an early sample or a scaled model of an early product which
is intended to build in order to test a concept and then be replicated in larger scale.
Thus it can be re-designed if necessary before the product is manufactured commercially.
From designers point of view prototyping is some design of the first order meaning just
trying to come up with an idea and finally to come up with the part. It is like to be
able to put everything together. They check whether the idea is going to work or not.
Furthermore after building the prototype we can take it to the laboratories and probe
it for its functional, thermal or electrical properties. We can measure it to make sure
everything is where it should be and of the right size and then it will fit or not if we put
it in the final enclosure.
Pay attention that there is a limitation for building an appropriate prototype that it is
just one instance of the hundreds of thousands of real product since each one of those
prototypes might have a slight difference due to the manufacturing and component
tolerances. So the question is, how many prototypes do you build? Everybody has to at
least build one prototype. Across most people build two, three or even four. Thus what
we face with is a large number of prototypes and that is the problem indeed because of

Chapter 1. Introduction

Figure 1.5: A scale model of an airplane in a wind tunnel for testing.

the high additional costs, depending on your type of products, a prototype could cost
many thousands of euros depending upon the component that we are putting inside our
model.
There is also a price to pay in terms of the schedule. Every time we build a prototype
it is going to take two, three or four weeks that introduce additional delay into your
development cycle. So what causes these iterations of the prototype? In fact in can
dome from many area including electrical problems such as signal and pair integrity or
EMC (Electromagnetic compatibility) for printed circuit boards. It could be caused by
a functional problem such as hooking up with the scheme or plan incorrectly. It could be
a manufacturing, reliability or a fit issue. So imagine that your prototype is an assembly
of different parts and it is going to be implemented in an enclosure but it does not fit
properly. It might clash with some other parts. So what can we do to eliminate all these
iterations of making prototypes?
Virtual Prototyping
A prototype is a digital model that contains links to all data, information and codified
knowledge required to design, engineer and manufacture an artifact [6].
So a virtual prototype provides a lot more flexibility with respect to the physical ones.
We can simply check the range of tolerances we spoke about earlier. In summary it is
important that we have only one instance of model with physical prototype instead an
infinite number of instances with virtual prototyping technology.
A real benefit of virtual prototyping is that the iterations are cheap. So we can very
quickly make a change to the design and re-analyze it. With virtual prototyping there

Chapter 1. Introduction

could be multiple engineers all working on the same virtual prototype at the same time,
performing different types of analysis. So virtual prototyping can eliminate real physical
prototypes which resulting savings in development cycle time and cost.
It is considered as a business challenge since it is a trade-off between cost and performance and on the meanwhile it would be a challenge to the design engineers too. The
designer is always in the trench and thinks of how to make a compromise between cost
and performance meaning he has to think what is giving the most acceptable performance with the lowest cost.
It is all about how globally companies will turn into competitive business. If there is
lower cost solution for a design and someone around the world is going to find it then
it means that the solution will not be for the rest of the companies. In other words, a
business firm prospers if it is the leader in finding the solution at earlier time.
The reports of Aberdeen group [7] shows that the companies or manufacturers that
make an extensive use of the simulation at the early stage of the product design process,
surpass other companies because of earlier release of their products either complex or
simple ones, the cost of their product and the quality targets. So the best in class
manufacturers hit the revenue with $1.9 million lower costs for intricate products and
$21,000 for simple products
After the prototyping is done, an idea has been proved. Then the product is adapted to
go into making a large number of same parts. Take note that the prototype has to get
to the finished product to give the visual impact required to the customer and actually
it is what a customer is looking for.
Nowadays companies are much more looking for what called a Low-hanging fruit. A lowhanging fruit is simply anything that adds to the product and makes it more attractive.
It attracts more the customer, the end-user or the purchaser. Sometimes it is simple
things that can draw the product to the next level in terms of sale or in terms of usability
through their look for simple things.
Computer Simulation
Computer simulation has become an inseparable part of mathematical modeling and also
many other natural systems such as what astronomy, chemistry or physics. Basically a
virtual model is created by some analytical tools and is run in order to have an estimation
of the performance of the too complex systems. The model benefits from algorithms
and equations that used to capture the behavior of the system being modeled while on
the contrary a simulation has already got the assumed equations and by running the
simulation the user has the advantage to check the results.

Chapter 1. Introduction

10

Figure 1.6: Complete computer simulation of a bacterium

As a matter of fact it helps us to obtain a more accurate prediction of the systems

performance which respects the other side of the trade-off meaning the costs.

1.4.1

LMS AMESim

AMESim is a commercial simulation software for the modeling and analysis of multidomain systems. It is part of systems engineering domain and is located in the Mechatronic category.
The software package consists of tools used to model, analyze and predict the performance of mechatronics systems by using electrical and mechanical block diagrams which
are of utmost importance. Models are described using nonlinear time-dependent analytical equations that represent the systems hydraulic, pneumatic, thermal, electric or
mechanical behavior. This approach provides the capability to simulate the behavior of
systems, hence it is used earlier in the system design cycle.
In the next chapter this software is used to simulate and optimize the mechanism of a
hinge.

Chapter 2

Simulation
2.1

Case Study

The Hotpoint Electric Heating Company (generally known simply as Hotpoint) is a

British brand of domestic appliances. The brand is currently fully owned by Indesit,
an Italian multinational company based in Fabriano, Ancona province, Italy, a leading
appliance manufacturer in Europe. In this project we concentrate on the oven of Indesit
Company, namely LUCE SX 1046LPX S where its picture is shown in Fig 2.1.
An oven is an insulated container used to heat or bake food. There are different types of
ovens like Earth oven, Ceramic oven, Gas oven, Mansonry oven and Microwave oven.
More recently oven has become slightly more high-tech in terms of cooking strategies
and functions.
The features and technical data of the case study are shown here.

Figure 2.1: LUCE SX 1046LPX S

11

Chapter 2. Simulation

12

Figure 2.2: Features

Figure 2.3: Technical details

We want to analyze the opening and closing mechanism of the door of the oven. Thus
in the next section we are going to decompose the door in order to obtain a better
understanding of the movement and associated compartments.

2.1.1

Computer Aided Design Drawings

Computer-aided design (CAD) is the use of computer systems to assist in the creation,
modification, analysis, or optimization of a design [8]. CAD software benefits the designer in sense of productivity, promote the quality of design since every operation could
be used in the software and there is no need to draw an item on paper, enhance transferring through documentation, and to create a database for manufacturing [9]. A CAD
file could get any suffix for print, machining, or other manufacturing operations. In
mechanical design, it is also known as computer-aided drafting (CAD) which describes

Chapter 2. Simulation

13

Figure 2.4: Housing

the process of creating a technical drawing on a standard sheet with the use of computer
software such as SolidWorks or CATIA.
CAD software for mechanical design uses either vector based graphics which is the
use of the simplest geometrical objects such as lines, circles, curves, polygons, arcs,
etc. to illustrate the objects of traditional drafting, or raster graphics demonstrating
the overall appearance of designed objects involving more than just shapes. Like in
the manual drafting of technical and engineering drawings, a CAD file has to convey
information, such as materials, processes, dimensions, and tolerances, according to the
specific application used.
As explained earlier, CAD could be used either to design 2D drawings such as sketches
and 2 dimensional figures using basic geometrical properties like curves or to design
solids in three dimensional (3D) spaces.
In our case study, The oven is created by assembling different parts like exterior shells,
side panels, door, gasket and etc. which is created separately, then assembled to form
the final product. Each part like door may include different sub-parts itself. Each part
is modeled in the CAD software Solidworks. Generally the parts are divided into two
categories. The ones used in the base and those in the door. The drawings of the parts
are represented in here to have a better evaluation of the model.
1- Base Parts
First, the housing also known as coving is a piece of material with different shapes used
to cover transitions between surfaces from solid milled plaster (Figure 2.4)

Chapter 2. Simulation

14

(a) Interior shell

(b) Interior middle shell

(c) Back shell

Figure 2.5: Shells to be placed on the back of the oven

Besides, there are Shells molded to fix the housing to the wall from behind for modern
classic looking kitchens. The oven we deal with is considered as a pyrolitic oven. Using
high temperatures of 500 C, a pyrolytic oven turns food residue and grime into ash which
can then be easily wiped clean using a damp cloth. No harsh chemicals or scrubbing
required! For safety during use, the pyrolytic control securely locks the oven door, and
only releases the lock once the temperature drops below 300 C. Because pyrolytic ovens
are well insulated, efficiency is assured. Self-cleaning ovens usually have more insulation
than standard ovens to reduce the possibility of fire. The insulation also reduces the
amount of energy needed for normal cooking. Furthermore, these ovens have walls
coated with heat and acid-resistant porcelain enamel which is a material made by fusing
powdered glass to a substrate by firing.
This is a Convection oven known either as a fan-assisted oven or a fan oven that has
fans to circulate air around food on the contrary of Conventional oven where no fan is
used inside the oven. Then the heat generated inside the oven is transferred into the
ambient by side sheets implemented on the left and right side of the housing.

Chapter 2. Simulation

15

Figure 2.6: Side panels

Figure 2.7: Gasket

A Gasket act as a mechanical sealing between two surfaces in order to prevent the
leakage. Gaskets allow less-than-perfect mating surfaces on machine parts where they
can fill irregularities. Gaskets are commonly produced by cutting from sheet materials
such as silicons, fiber glass, copper and etc.
2- Rotating Parts
The rotating part includes the door which is composed by several portions like a steel
sheet, pivots, rollers, condensation collector, handle and etc.
A hinge is a mechanical device that connects two solid objects, allowing them to rotate
with a limited angle with respect to each other around a fixed rotation axis.
The opening and closing mechanism of the oven is based on the hinge parts which is
explained in detail here. In the other words, the hinges aid the door of the oven to open

Chapter 2. Simulation

16

(a) Condensation collector

(b) Handle

(c) Exterior Design

(d) Steel Sheet

(e) Facade Part 1

(f) Facade Part 2

Figure 2.8: Door compartments

Chapter 2. Simulation

17

(a) Door frame

(b) The Slider placed in lateral side of the door

Figure 2.9: Door frame

(a) Hinge side view

(b) Hinge Front view

Figure 2.10: Diplomat hinge

and close. It includes many elements inside such as springs, a roller, a main stay rod
and etc. Besides, there is a cam profile mechanism aids the door snap back to its initial
position.
The hinge of this case study is called Diplomat that is used mainly in ovens and is one
of the last generation of its type.
Of utmost importance is the hinges location where is fixed inside the base at one end
and is rotating at the other end.

Chapter 2. Simulation

18

Figure 2.11: The location of hinge

2.2

AMESim

In the first chapter the software LMS AMESim was introduced briefly. This software is
used to model a mechanism composed of different parts. To create a simulation model
for a system, a set of libraries is used, like signal library, mechanical library and etc.
They contain pre-defined components each containing different mathematical equation
for different physical domains. The elements in the system have to be connected and
for this purpose each icon has ports, which have several inputs and outputs. Causality
is imposed by connecting the inputs of one icon to the outputs of another icon (and
vice-versa).
AMESim libraries are written in C language and it also supports Modelica which is a
free source, object-oriented, equation based language to model complex physical systems
containing, e.g., mechanical, electrical, electronic, hydraulic, thermal, control, electric
power or process-oriented subcomponents [10].
The modelling of a system is done in four steps:
1. sketch mode: in which the different components are linked.
2. submodel mode: in which the physical submodel associated to each component is
chosen.

Chapter 2. Simulation

19

Figure 2.12: AMESim Simulink lab

3. parameter mode: in which the parameters for each submodel are set.
4. run mode: in which the simulation is started and results analyzed.
These four steps are discussed later. Between the submodel and parameter mode, the
AMESim model is compiled. Under the Windows platform, LMS Imagine.Lab AMESim
works with the free Intel compiler , which is provided with the software.

2.3

Opening-closing mechanism of the door of the oven

The opening-closing mechanism of the hinge is shown in fig.2.13. The whole mechanism
is divided into six different parts. This separation is done in order to clarify the function
of each sub-system.
Main body, stay-rod
The main body has seven ports connecting different parts together. We use relative
coordinates reference to make it easier to calculate the x and y coordinates of ports one
to seven. The absolute angular position of the main body is equal to 90 in Cartesian
coordinates. The weight of the hinge is considered as 1kg and the moment of inertia as
0.01 kg.m2 .
The main body is demonstrated in fig.2.10.b and fig.2.14.

Chapter 2. Simulation

Figure 2.13: Opening-closing mechanism of the door of the oven

20

Chapter 2. Simulation

21

Figure 2.14: Main body block diagram

Figure 2.15: Port positions in parameter mode

Figure 2.16: Real hinge

Chapter 2. Simulation

22

Figure 2.17: Force directions

The first part is about the input force. We apply an amount of force to the handle of the
oven, as we hold and move it. The value of force starts from 0N and reaches a maximum
(the summit) in a fraction of second, then suddenly falls back to the value 0N. So it is
a parabolic function and similar to a Gaussians distribution function. The experiments
in the next chapter confirm this hypothesis.
An amount of force 100 N is applied both in the x and y directions of the oven fig.2.17.
As explained before, the force can be considered as a function of Gaussians distribution
form. The function of this type has the formula

f (x) = a1 exp(((x b1 )/c1 )2 )

The parameter a is the maximum force applied to the handle and b is corresponding
time of the peak. c is about how steep the parabola should be considered. So we choose
a as 100N, b and c as 1.
As illustrated in fig.2.18, a constant gain is used to insert the value of maximum force,
A clock whose input is the duration of the simulation and a function block diagram to

Chapter 2. Simulation

23

(a) Block diagrams related to the first part

(b) Gaussian distribution of the force

Figure 2.18: First part of block diagrams

insert the formula of Gaussian distribution function. Besides, there is gain to invert the
direction of the force with respect to y axis.
A velocity sensor and an acceleration sensor are implemented to have an idea about
the outcome velocity and acceleration where is used usefullly in the next chapter for
optimization of the mechanism.
Second part includes a reference fixed body and a driver revolute pivot. In addition
there is a rotary Coulomb friction block diagram.
The reference fixed body is demonstrated in fig.2.10.a. It is located in absolute Cartesian
coordinate (0, 0) and the corresponding absolute angular position 0 .
The driven revolute pair is used to connect the reference fixed body to the main body by
a pivot. Apparently there exists a friction force since two metallic surfaces rotate with
respect to each other. This friction is modeled with a rotary Coulomb friction which is
considered as dry friction. Thus the force needed to move the mechanism must overcome
the static friction and move the mechanism.
Coulomb described the friction force as:
F = Fc .sign(v)
Where the friction force Fc (Coulomb friction force) is:
Fc = Fn
Note that Fn is the normal force and Fc is tangential force.
The dynamic friction coefficient is set to 0.02 and the maximum Coulomb torque as 30
Nm. Rotary stick velocity threshold is set as default and the diameter on which the
friction acts is 10 mm.

Chapter 2. Simulation

24

There is also a constant gain whose value is multiplied by the maximum Coulomb torque
to calculate the dynamic friction force. The value of the constant gain must be a
fraction lower than 1.
In third part, a tensile spring with its linear Coulomb friction is modeled. The spring
tensile force is calculated manually according to the formula:

k=

G.d4
8.D3 .na

(2.1)
(2.2)

Where d is the spring diameter equal to 1 mm, G is shear modulus and 80 Gpa for steels,
na is the number of active coils equal to 20 and D the nominal diameter is 15 mm. Then
k would be around 5000 N/m and the amount of preload F0 is estimated at 300 N. The
signal input is 3 meaning the friction force is obtained by the constant signal equal to
10 N. The maximum Coulomb friction force is equal to 5 N and the dynamic friction
coefficient is dimensionless and equal to 0.3.
In reality the door of the oven does not rotate backwards when it is in closed position.
In order to make the software understand this concept, two springs and dampers with
very high stiffness are implemented behind the door. In fact, it is called a wall because
it restricts the rotation of the door of the oven when the angular position of the door is
above 90 .
To do so we can put the stiffness as a function of the position, using a step block (red
block section). The equations in the functions ((tanh((x) 100000) + 1)/2 1000000)
are set in a way to deliver a high damping ratio if the condition is not satisfied.
In fifth part a three-port body is observed where in reality is the part where the follower
pushes in order to compress the shorter spring. In parameter mode, we used a relative
coordinate system with 90 initial absolute angular position. The x and y position of
the initial absolute position is 15 mm and 3 mm respectively.
There is a compressive spring whose high stiffness is calculated according to the formula
2.1 and is evaluated around 10,000 N/m and spring force with both displacements zero is
estimated at -500 N. The other important element in this section is the cam and follower
mechanism. Pay attention that the cam is fixed here and does not move. In fact there
is a special type of cam and follower mechanism. The profile of the cam is essential in
order to be put as an input in parameter model fig.2.22. It is extracted from the CAD
drawings and is shown in the figure 2.23.

Chapter 2. Simulation

25

(a) Second part of block diagrams

(b) Corresponding part in real hinge

Figure 2.19: Second part of block diagrams

Chapter 2. Simulation

26

(a) Third part of block diagrams

(b) Corresponding part in real hinge

Figure 2.20: Third part of block diagrams

Chapter 2. Simulation

27

(a) Fourth part of block diagrams

(b) The graph of tanh funcion

Figure 2.21: Fourth part of block diagrams

Figure 2.22: Set parameters for cam and follower mechanism

Chapter 2. Simulation

28

Figure 2.23: Cam profile shown with light blue line

Figure 2.24: Cam profile in 360

Then the characteristic equation of the cam profile is imported in the software MATLAB
in order to change the data from an equation to the numerous points Cartesian coordinates. Because AMESim table accepts the cam profile in terms of its characteristic
points and not characteristic equation. Then cam profile has to be converted into the
scale of 360 . The angular displacement of the cam is put equal to 29.61 and the rest
of the data is represented in the fig 2.24.
The angle indicating the angular displacement of the cam is calculated according to
the formula below that calculates the angle between two vector a and b.

Chapter 2. Simulation

29

cos =

a
.b
abs(a).abs(b)

there are two points at both ends of the cam profile that creates two vectors with center
of the cam. the positions of these two points are (11.19, 18.29) and (18.37, 11.19).
Furthermore there is a signal from velocity sensor in the main body where enters the
angular velocity of the cam. There exists a force transferred from the second body into
the main body, thus a force sensor is implemented to transfer the reaction of the spring
to the main body (stay-rod).
In sixth part an elastic end stop is used so that the door of the oven is not allowed to
rotate into an angle less than 0 or in a better word the door rotates maximum 90 .
The required parameter of the elastic end stop is shown in fig.2.26.b.
In Sketch mode we designed the system with block diagrams, then in Submodel mode
every component in the system must be associated with a mathematical model. This is
a collection of mathematical equations and their implementation as a piece of computer
code.
The AMESim nomenclature is to describe the mathematical model of a component of
a system as a submodel. The term model is reserved for the mathematical model of
the complete system. AMESim contains a large collection of submodels. As far as it is
appropriate, the process of associating submodels with components is automated.
So, we click on the submodel mode button in the toolbar. With AMESim a single
component may have more than one submodel that can be associated with it. For
example for Dynamic friction model, we have many submodels and we chose the Coulomb
dynamic friction model using tanh function. We simply use Premier submodel facility
for the rest of the model.
Next setp is setting parameters. AMESim performs various checks and creates an executable code for your system. The System Compilation window appears giving you
some technical information about the equations it must solve to perform a simulation.
We changed it as explained before.
For Running a simulation there are some steps. First, we should enter the Simulation
mode, then we set the Run parameters. A dialog box appears (fig.2.28) that allows you
to change the characteristics of a run like start time and final time. The dialog box
is comprised of two tabs General and Standard options. The settings are set to
default.
Then we click on Start a simulation, Then run monitor dialog box appears.

Chapter 2. Simulation

30

(a) Fifth part of block diagrams

(b) Corresponding part in real hinge

Figure 2.25: Fifth part of block diagrams

Chapter 2. Simulation

31

(a) Sixth section blocks

(b) Elastic endstop parameter

Figure 2.26: Sixth section Block diagrams

Figure 2.27: AMESim parameter

After running the simulation, it is the time to import the CAD models into AMESim,
assembly mode. Thus we double click on the assembly icon on the left upper corner of
the graphic window and this window appears fig.2.30.
Above the CAD model, a bar is shown (fig.2.31) where controls the animation part of
the assembly according to the parameters selected for the models. With this bar, there
is the possibility to play the animation of the oven with Large opening meaning without
snaping back. The picures are illustrated here fig.2.32.
The movement characteristics of the door including its position, velocity and acceleration
are listed in the variable list after running the simulation.
The graph for absolute angular position, absolute angular velocity and absolute angular
acceleration is plotted here.

Chapter 2. Simulation

32

Figure 2.28: Run Parameters

Figure 2.29: Start a simulation

As can be seen clearly in fig.2.34, the angular position is at 90 in its initial position,
then step by step it decreases until it reaches a minimum point and after that a slight rise
of the graph occurs where comes from the reaction of the springs. In absolute velocity
graph the velocity is increasing until it reaches a maximum point (around -32 rev/min)
then it tends to decrease but again it increases a little bit due to the reaction of the cam
and follower mechanism and the compressive spring and tensile spring. Finally it ends
with 0rev/min where it the door stops moving. For acceleration as it is observed, there
is an initial acceleration followed by a sudden decrease to the value 0.
There would exist a second condition too where the door snaps back. That is because

Chapter 2. Simulation

33

Figure 2.30: Assembly mode

Figure 2.31: Assembly mode

(a) Door in closed position(initial position)

(c) Door in opening mode 2

(b) Door in opening mode 1

(d) Final door position

Figure 2.32: Animation of the opening-closing mechanism of the door of the oven
with Large opening

Chapter 2. Simulation

34

Figure 2.33: Stay-rod variable list

(a) Body angular postion

(c) Body angular acceleration

(b) Body angular velocity

(d) Body angular velocity with its minimum point

Figure 2.34: Graph for position, velocity and acceleration of the main body

Chapter 2. Simulation

(a) Door in closed position(initial position)

(c) Door in opening mode 2

35

(b) Door in opening mode 1

(d) Final door position

Figure 2.35: Animation of the opening-closing mechanism of the door of the oven
with small opening

the force required to drive the door is not enough and the door opens maximum up
to 20 to 25 and no more. To find the value of the force, a batch run has to be set
and run in order to check when the door comes back to its original position and then
check the corresponding force where in this case study is 38N. So with the help of the
new generation of Diplomat hinges it would be possible that the door travels the same
distance backward thanks to the cam and follower mechanism and two tensile and compressive springs where apparently are preloaded (One stretched and one compressed).
Then the graph for absolute angular position, absolute angular velocity and absolute
angular acceleration varies according to the movement of the door where the animated
picture is shown here.
As can be seen in fig.2.36, apparently the initial angular position is at 90 . It decreases
rapidly as time passes and reaches to the value 62.9 . Then it starts rotating upward.
This reaction as explained before is due to the preloaded tensile and compressive springs
and the cam and follower mechanism. So it tends to go back to 90 again. The parabola
of velocity is exactly as expected before. There exists a rise and a fall that indicates the
opening and closing of the door respectively. If we take a closer look at the velocity-time
diagram, we will notice that the positive area above the time axis and the negative area
below the time axis are the same. Indeed it means that the system does a symmetric
movement. The magnitude of both peaks of the velocity diagram is the same (about
16rev/min). For acceleration diagram the situation is the same. The system profits

Chapter 2. Simulation

(a) Body angular postion with snap back effect

(c) Body angular acceleration with snap back effect

36

(b) Body angular velocity with snap back effect

(d) Body angular position with its minimum point

Figure 2.36: Graph for position, velocity and acceleration of the main body with
snap back effect

from an initial large acceleration followed by a steep fall.

In the next chapter, optimization of the oven will be discussed in detail.

Chapter 3

Optimization
3.1

Experiment

In first section of chapter three, it is being shown how to measure angular acceleration,
angular velocity and load. The test starts with attaching three different sensors including
the load sensor, the gyroscope and the accelerometer sensor to the different parts of the
door of the oven. The sensors and the way they are attached to the oven are shown in
fig.3.1.
The load sensor Model LTH300 is a Donut Load Cell built for Press or Inline compression
applications.
We used a tri-axial accelerometer with sensitivity 5.1 mv/(m/s2 ), weigh 100 g, and a
frequency range between 1 to 4000 Hz and with a titanium housing. Adhesive mounting is often used for temporary installation or when the test object surface cannot be
adequately prepared for stud mounting.
The Vibrating Structure Gyro is a solid state rate sensor that is designed around a
piezoceramic cup that can measure coriolis forces. This allows the VSG to produce a
highly accurate output that is proportional to the rate of turn. Electronics for filtering,
control, drive and demodulation are all packaged along with a single axis sensing element
inside a small package. Wide ranges of input rates (up to 3000 per second) are available
[11].
After installing the sensors, we start out the experiment by rotating the door of the oven
from closed position to open position (inclined position). Thus we activate all the sensors
simultaneously and start doing measurements for load, angular acceleration and angular
velocity. Note that the output of the sensors is in mv that must be converted into v,
37

Chapter 2. Optimization

38

Figure 3.1: An overview of the sensors installed on the oven

(a) FUTEK load sensor

(b) Load sensor structure

Figure 3.2: Load sensor

Chapter 2. Optimization

39

(a) 3D view of Accelerometer

(b) Accelerometer structure

Figure 3.3: PCB Pieszotronics accelerometer

(a) 3D view of gyroscope

(b) Gyroscope drawing

Figure 3.4: Single Axis Watson VSG

then multiplied by a coefficient in order to change the unit from Volt to degree/s2 , N
and degree/second for acceleration, load and angular velocity respectively.
Measurements are stored in text files through the softwareLabview NI. The characteristics of the text file including four columns introducing time, acceleration, angular velocity
and load respectively. The result of the one-time measurement is shown in fig.3.6.
The curve data are fitted because of presence of the noise. The parabolic data curves are
fitted for angular acceleration, load and angular velocity. To do so, we use the software
MATLAB and the function cftool to fit the curves.
Many functions like Polynomial (up to 19th order), interpolant function, Gaussian function, Power function, Rational function and etc. are tried. Gaussian curve fitting gives
the best result figs.3.7., 3.8. and 3.9.

Chapter 2. Optimization

40

(a) Open postion

(b) Closed position

Figure 3.5: Initial and final position of the door of the oven

Figure 3.6: Experiment data

Chapter 2. Optimization

41

Figure 3.7: Curve fitting of load data vs. time

Figure 3.8: Curve fitting of acceleration data vs. time

In the next section it will be explained in detail how to implement the result of experiments in AMESim software and utilize the optimization package for parameterization
of the system.

3.2

Optimization and parameterization of the case study

The load function is implemented in the parameter mode of the function block diagram.
As we already know, the load is considered as an input.

Chapter 2. Optimization

Figure 3.9: Curve fitting of velocity data vs. time

Figure 3.10: New load function f (x) = 356.8 exp(((x 3.691)/0.07581)2 )

Figure 3.11: New load function f (x) = 356.8 exp(((x 3.691)/0.07581)2 )

42

Chapter 2. Optimization

43

Figure 3.12: New load function f (x) = yexp(((x3.777)/0.1385)2 )+zexp(((x

3.68)/0.07037)2 ))

Thus the first section of the system is modified properly. It means that the load function is inserted in the parameter mode of function block diagram. The velocity or
acceleration of the model has to be compared with the velocity or acceleration values
of the measurement performed in the previous section. So another section has been
added to the model to insert the velocity or acceleration of the real model obtained by
the measurement. We do this by selecting another set of block diagrams including a
function block diagram, a clock and a constant gain. Besides, a summing junction is
needed to compare the two results meaning the results of the model and the ones of the
measurement.
Take note that the goal of the optimization is based on minimizing the output of the
summing junction as close as possible to the value zero. Actually this is why summing
junction icon is placed in the modified system. Because the output of the summing
junction is the subtraction between the velocity (or acceleration) of the model with
velocity (or acceleration) of the measurement.

3.2.1

Export

In AMESim we talk about submodel parameters, global parameters and submodel variables. For the export module we talk about input and output parameters.
Input Parameters are values associated with a name that are directly or indirectly needed
by the simulation to run. Every submodel parameter visible in Parameter mode and
every global parameter of a model can be used as an export input parameter fig.3.15.

Chapter 2. Optimization

44

Figure 3.13: Export setup

Output parameters are values associated with a name that are produced by the simulation or that are computed from other output parameters. Every submodel variable
could be used as an export output parameter. Such outputs are called Simple Output
Parameters fig.3.16.
Compound Output Parameters are post-treated outputs: they are computed from user
defined expressions which involve any export Input Parameters, Simple Output Parameters or other Compound Output Parameters. Certain functions may also be used fig.3.17
When using AMESim and other commercial design exploration/optimization software,
the Export facility is used first. Using this tool, it is easy to change the parameters of
the model and get post-treated results of the simulation. The Export facility sets up
parameters in a format suitable for AMEPilot.
Setting up the export
First we shift into Parameter mode. So an Export Parameters Setup dialog box is
created. Note that there are three tabs. Input Parameters section is currently selected.

Set the parameters needed for the export

The aim of optimization is to modify the values of every element used in the hinge.
So stiffness of both springs and amount of springs preload besides the friction force in
different parts of the hinge are selected as input fig.3.15. The names of the selected
parameters are changed according to their functionality.

Chapter 2. Optimization

45

Figure 3.14: Export setup

Figure 3.15: Input parameters

By default the value types are Real values. The export parameters will be accessible
outside of AMESim. As said before we edited the Export Name and Default value fields
as what we calculated in the second chapter. We switch to the Simple output parameters.
Put the output from the summing junction in this section and change the name into
Dif in vel3.16.
The aim of optimization is to approach two different systems meaning the real system
and the modeled system. Thus several criteria could be used in order to be posed inside
the system. We have to minimize either:

1. Maximum difference between the velocity (or acceleration) of the real system and
of the model,

Chapter 2. Optimization

46

Figure 3.16: Simple output parameters

2. Maximum squared difference between the velocity (or acceleration) of the real
system and of the model,
3. Final value of the integral of absolute difference between acceleration (or speed),
or
4. Final value of the integral of squared difference between acceleration (or speed).
Finally we change the tab to Compound output parameters and eight expressions are
formulated according to the optimization criteria mentioned before. Four of them belong
to optimization through velocity criteria and the other four belongs to optimization
through the acceleration criteria fig.3.18.
In the next step we use the optimization tool in the software LMS AMESim.
Therefore we enter the design exploration by Select Analysis and then Design exploration
or click on the Design exploration button

to produce the Design Exploration dialog

box.
Then we right-click on Optimization and click on New study and then in the new window
Design exploration definition appears. We leave the lower bound and upper band as
what defined by default. Because as explained earlier in the overview, the state of the
art consists in first performing a global optimization (genetic algorithm) in order to
explore the design space and to avoid local minima, and then to extend the study with
a local algorithm (NLPQL) to refine the convergence.
At the end of optimization process, the optimized input parameters are given in the
table 3.1. The best solution found is through the fourth criterion which is minimizing

Chapter 2. Optimization

47

Figure 3.17: Compound output parameters

Figure 3.18: Simple output parameters

Figure 3.19: Design exploration dialogue box

Chapter 2. Optimization

48

Figure 3.20: Genetic Algorithm

(a) Optimization input

(b) Optimization output

Figure 3.21: First optimization based on the Genetic Algorithm

Chapter 2. Optimization

49

Figure 3.22: Solution progress dialogue box

Table 3.1: Best solution found via the criterion minimizing final value of the integral of squared difference between speed of the modeled system and that of the real
experiment.
Compressive spring stiffness
Tensile spring stiffness
Maximum Coulomb friction force for compressive spring
Maximum Coulomb friction force for compressive spring
Maximum Coulomb friction force for ground pivot
Fraction of maximum friction force for ground pivot (between o and 1)
Friction force for the tensile spring
Fraction of maximum friction force for compressive spring (between o and 1)
Tensile spring preload
Compressive spring preload
Optimization error

2.61508737392910e+003
2.09336993819491e+003
1.11562549751767e+001
7.70230749470754e+001
2.56671657452170e+000
1.28638611059235e-001
3.39469034866910e+001
3.58883552616509e-001
5.82768015744605e+001
-6.76591341123854e+001
3.75098995130593e-001

final value of the integral of squared difference between speeds of the real model and
the simulated one. Because the optimization error is no more than 0.38% according to
the table 3.1 while in other criteria the error exceeds 5%.
The velocity graphs for the real hinge and modeled hinge are illustrated here in fig.3.23.

Chapter 2. Optimization

50

Figure 3.23: Optimized velocity of the model vs. Real velocity obtained by measurement

3.3

A Model of the User as a Mechanical Load

At the beginning of this section the concept mechanical impedance is explained clearly
because the following models presented here are based on mechanical impedance of
the user. Impedance is nothing rather than a resistance or stiffness in materials. The
materials that are soft have lower impedance while the more stiff ones have higher
impedance. Humans mechanical impedance depends greatly on the type of grasp thus
three categories of grasps are defined here.

1. contact grasp which is touching an object with one finger or even with the whole
palm of the hand,
2. precision grasp which is holding an object in a precise way. It is usually done with
two fingers one thumb and the other acts as a holder,
3. power grasp which is holding an object tightly and is normally done with whole
fingers in the sense that thumb is used as a counter bearing and the rest of the
fingers to hold a knob or a handle firmly.

3.3.1

Model

KERN [12] defined an eight-element model fig.3.24 for measuring the impendace. The
model can be classified in three different groups used for many grasping situations: Z 3

Chapter 2. Optimization

51

Figure 3.24: Eight-element model of the users impedance, modeling the passive mechanics for frequencies >20 Hz.

models the elasticity and damping of the skin that is in direct contact with an object
(handle in this case study). Z 1 is the central element of the model and depicts the
mechanical properties of the influencing body parts - normally fingers. Z 2 gives an
insight into the mechanical properties of the limbs, usually hands. This model also lets
up to assume an amount of preload in the joints.
The Kerns electrical board is used to obtain the generated output value F out which
The eight-element model of the users impedance proposed by kern is modeled in the
AMESim for the first time. In fact several block diagrams are used to simulate the
Kerns electrical circuit. Thus the block diagrams listed here are connected in fig.3.26 in
AMESim.
1. Clock in order to simulate the time,
2. The function block diagram to insert the formula of humans impedance transfer
function which will be discussed in section impedance measurement parameters,
3. A division block diagram,
4. A splitter junction to share the force exerted in both x and y directions of the
handle.

Chapter 2. Optimization

52

s2 m2 + k1 + d1 s
s
s
1 1
+
)
Z2 = (
d2 s + k2 sm1
d3 s + k3
Z3 =
s
Z B = Z1 + Z2
Z1 =

(3.1)
(3.2)
(3.3)
(3.4)

According to the transfer function Z H and the formula

v spo
F out
v v ind
= out
F out

GH3 =

(3.5)
(3.6)

=YH

(3.7)

(3.8)

1
ZH

(3.9)

The generated output value F out is calculated by multiplying the velocity by the mechanical impedance according to the equations 3.5 and 3.9. The velocity is obtained
by getting the signal from the velocity sensor which was implemented in the first
part of the model and the mechanical impedance is measured by performing a measurement. The value of the mechanical impedance will be obtained from section Impedance
measurement parameters.
This model could be substituted by the real or evaluated load parameter implemented
in the second chapter by the use of signal library and mechanical library for conveying
the force to the input system.
Impedance measurement parameters
The acquisition of mechanical impedance is a daily problem. The main principle to
acquire mechanical impedance is to apply an excitation to the system and measure the
corresponding output force and velocity at the same time.
KERN[13] performed the impedance measurements based on a sinoid-sweep frequencies.
The excitation is made with a specified force of 2N amplitude at the sensor and accordingly the mechanical impedance of the sensor is measured fig.3.30. The following table
3.29 provides the parameters for the model illustrated in fig.3.24.

Chapter 2. Optimization

53

Figure 3.25: Global parameter setup dialog box

Figure 3.26: Hand user elements implemented in AMESim

ZH = Z3 k ZB

(3.11)

=(

(3.10)

s
s2 m2 + k1 + d1 s 
1 1 1 1
+
+
) ) )
s
d2 s + k2 sm1

(3.12)

In Parameter model fig.3.28 the function is based on the equation of mechanical impedance
acquired from equ.3.12.
These parameters are shown here in fig.3.29 are transferred into the Global parameter
setup dialog box fig.3.25.
The plot of the force for Power grasp meaning holding an object woth at least one
finger and a counter-bearing is shown in fig.3.31.

Chapter 2. Optimization

54

Results
According to the fig.3.31. the generated output value F out is about 550N which is the
force along the tissues including the limbs, muscles and etc.
Based on the results obtained in the second chapter the maximum force applied is about
250N. If we make a comparison between these two results, it is being found that the
internal force along the tissues of bodys hand is about 2.2 times more than the external
force applied to the handle of oven. It means that as estimation half of the force is
damped when it exerts to the handle. It could be due to the numerous bones, muscles
and limbs which are implemented in hand membrane. The hand itself contains 27 bones
and 33 muscles. Also we can think of action potential [14] of systems physiology that
is a short-lasting event in which the electrical membrane potential of a cell rapidly rises
and falls following a consistent trajectory.

Chapter 2. Optimization

Figure 3.27: Hand user elements implemented in main model

55

Chapter 2. Optimization

Figure 3.28: The mechanical imedance function in parameter mode

Figure 3.29: Mean values of the mechanical impedance model according to figure 3.24
for different grasping siutations

Figure 3.30: Impedance measurement settings for different grasps [13]

56

Chapter 2. Optimization

Figure 3.31: The plot of the force exchanged between finger-tip, bones and muscles

57

Chapter 4

Conclusion
The new technology Virtual Prototyping as its name offers is about building virtual
prototypes. This method challenges what was believed as traditional product design
process in the sense that the demanded model or needed system could be simulated by
means of software and several different analyses could perform on the virtual model.
The case study is about simulation and optimization of an oven hinge through the
software LMS AMESim. After understanding how the mechanism works, the process
of its simulation starts. Icons belonging to different categories are connected though
their ports to simulate the movement of the hinge in detail. The proper parameters
are calculated for each block diagram like the stiffness of the compressive and tensile
springs. Then we let it move into running the simulation. In the next step the CAD
models are imported from the software Solidworks into AMESim to have a 3D perception
of the mechanism assembly and how it moves. Optimization via a combination of genetic
algorithm and NLPQL methods is done based on the criterion minimizing final value
of the integral of squared difference between speeds of the real model and the simulated
model. So an experiment is done to find out the real values of the input force and the
corresponding outputs velocity and acceleration.
After simulation we have access to the data regarding the acceleration, velocity and
position of the door at every moment. Furthermore we can read the information regarding the followers displacement and the associated force necessary to apply with the
aim of expected reaction of the door. After performing optimization, a collection of the
best data is gathered then the proper values of each component used in the system are
allocated. In the last part it is being illustrated that the force which is applied by the
hand of the user is in Gaussians shape verifying the experiment we did in the second
chapter.

58

Chapter 4. Conclusion

59

In the last part the modeling of a user as a mechanical load in sense of modeling the
hand of a user (in case of an oven it would be a power grasp) in AMESim thanks to the
KERNs eight element circuit and analogous of electrical and mechanical systems. Also
this field is expandable to other research field areas such as automobile industry. The
companies might be interested in simulate a virtual hybrid car and let it be driven in a
virtual path to be informed how many liters benzene or gasoline is needed to be fueled.
Nowadays companies are pushed into launching their products with the aim of first
lower cost process and second spending less time. In order to achieve this goal,
Virtual prototyping could be used as an interface among engineers and products.

Appendix A

References
[1] Tong C, Sriram R. Artificial intelligence in engineering design, vol.1. Academic Press;
1992 (Chap. Introduction)
[2] Birgit Verworn A structural equation model of the impact of the fuzzy front end
on the success of new product development Original Research Article Research Policy,
Volume 38, Issue 10, December 2009, Pages 1571-1581 Birgit Verworn
[3] Birgit Verworn Birgit Trends in customer service Aberdeen group November 2012
[4] Asimov, M introduction to design Prentice Hall, IEnglewood Cliffs, NJ (1962)
[5] Lawson, B How designers think Architectural Press, London
[6] Luciana DAdderio Crafting the virtual prototype: how rms integrate knowledge and
capabilities across organisational boundaries, Research Policy 30 (2001)
[7] Aberdeen Group (October 2006). Simulation-Driven Design Benchmark Report:
Getting It Right the First Time
[8] Narayan, K. Lalit (2008). Computer Aided Design and Manufacturing. New Delhi:
Prentice Hall of India. p. 3.
[9] Narayan, K. Lalit (2008). Computer Aided Design and Manufacturing. New Delhi:
Prentice Hall of India. p. 4
[10] Modelica and the Modelica Association
[11] Manual of Vibrating Structure Gyros
[12] THORSTEN A. KERN Haptic Engineering devices chapter 4. p. 68.
[13] THORSTEN A. KERN Haptic Engineering devices chapter 4. p. 66.
60

References References

61

[14] MacDonald PE, Rorsman P (February 2006). Oscillations, intercellular coupling,

and insulin secretion in pancreatic beta cells